LEARNING FROM EXAMPLES
Dr. J. Ujwala Rekha
Introduction
• Learning agents improve their behavior through
diligent study of their own experiences.
• Any component of an agent can be improved by
learning from data.
• The improvements, and the techniques used to make
them, depend on four major factors:
– Which component is to be improved
– What prior knowledge the agent already has
– What representation is used for the data and the
component
– What feedback is available to learn from
Components to be Learned
• A direct mapping from conditions on the current state to
actions
• A means to infer relevant properties of the world from the
percept sequence
• Information about the way the world evolves and about the
results of possible actions the agent can take
• Utility information indicating the desirability of world states
• Action-value information indicating the desirability of actions
• Goals that describe classes of states whose achievement
maximizes the agent’s utility
Representation and Prior Knowledge
• Representation of knowledge
– Propositional logic
– Predicate logic
– Bayesian networks
• Inductive Learning: when the output and examples
of the function are fed into the AI system, inductive
learning attempts to learn the function for new
data.
• Deductive Learning: going from a known general
rule to a new rule that is logically entailed
Feedback to Learn From
• There are three types of feedback that determine the
three main types of learning:
• Unsupervised Learning: the agent learns patterns in
the input even though no explicit feedback is supplied.
The most common unsupervised learning is clustering.
• Supervised Learning: the agent observes some
examples input-output pairs and learns a function that
maps from input to output
• Reinforcement Learning: the agent learns from a
series of reinforcements-rewards or punishments.
Supervised Learning
Supervised Learning
• When the output is one of a finite set of
values such as sunny, cloudy or rainy, the
learning problem is called classification
• It is called Boolean or binary classification if
there are only two values.
• When the output is a number such as
tomorrow’s temperature, the learning
problem is called regression.
Supervised Learning
• Figure 18.1 (a) shows some data with an exact fit by a straight
line 0.4x+3.
• Figure 18.1 (b) shows a high-degree polynomial that is also
consistent hypothesis because it agrees with all the data.
• This illustrates a fundamental problem in inductive learning:
how do we choose from among multiple consistent
hypotheses?
• According to Ockham’s razor, prefer the simplest hypothesis
consistent with the data.
• However, defining simplicity is not easy, but it seems clear that
a degree-1 polynomial is simpler than a degree 7 polynomial.
Supervised Learning
Learning Decision Trees
• A decision tree represents a function that takes as input a
vector of attribute values and returns a “decision”- a single
output value.
• The input and output values can be discrete or continuous.
• A decision tree reaches its decision by performing a sequence
of tests.
• Each internal node in the tree corresponds to a test of the
value of one of the input attributes and the branches from
the node are labeled with the possible values of the attribute.
• Each leaf node in the tree specifies a value to be returned by
the function.
Learning Decision Trees
Learning Decision Trees
Learning Decision Trees
Learning Decision Trees
• Over fitting happens when a model learns the detail and
noise in the training data to the extent that it negatively
impacts the performance of the model on new data.
• Over fitting becomes more likely as the hypothesis space
and the number of input attributes grows, and less likely
as we increase the number of training examples.
• For decision trees, a technique called decision tree
pruning combats over fitting.
• Pruning works by eliminating nodes that are not clearly
relevant.
T HANK YOU
